INDEX

Explanations

terms related to specific entities or brands, particularly those starting with "Re-" and followed by various suffixes

oai_token-act-pair · gpt-3.5-turbo

names and references to various entities and titles, particularly those related to specific works, characters, or discussions in media

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.7.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

»Ĵ

-0.76

ashtra

-0.72

ulhu

-0.68

wa

-0.64

ã

-0.63

æĪ¦

-0.62

acca

-0.60

Magikarp

-0.59

 Nadu

-0.57

é¾įå¥ĳå£«

-0.56

POSITIVE LOGITS

itory

0.88

earch

0.88

burse

0.83

otiation

0.80

issance

0.79

ndum

0.78

irmation

0.74

atives

0.73

inion

0.73

rocal

0.72

Activations Density 0.088%

terms related to specific entities or brands, particularly those starting with "Re-" and followed by various suffixes

names and references to various entities and titles, particularly those related to specific works, characters, or discussions in media

No Comments

No Known Activations