INDEX

Explanations

phrases related to a diverse range of topics and content, possibly related to interviews, discussions, or news articles

oai_token-act-pair · gpt-3.5-turbo

elements related to various discussions and topics in different contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

 Peck

-0.72

Corp

-0.69

BOX

-0.64

 Grac

-0.63

©¶æ

-0.61

Mog

-0.61

opter

-0.60

click

-0.60

 Ashe

-0.60

 Carnage

-0.59

POSITIVE LOGITS

 occasional

0.73

olis

0.70

 disclaim

0.70

azeera

0.70

 ones

0.69

 modifiers

0.66

cellaneous

0.65

 versa

0.65

ModLoader

0.64

ullivan

0.63

Activations Density 0.283%

phrases related to a diverse range of topics and content, possibly related to interviews, discussions, or news articles

elements related to various discussions and topics in different contexts

No Comments

No Known Activations