INDEX

Explanations

references to simulated or imitation versions of something

oai_token-act-pair · gpt-3.5-turbo

references to "mockumentaries" or similar formats

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

 Horizon

-0.74

 cryst

-0.72

ettel

-0.67

 violet

-0.66

pins

-0.64

ItemThumbnailImage

-0.64

OA

-0.63

 arrang

-0.63

omen

-0.63

 compan

-0.62

POSITIVE LOGITS

 Mock

1.00

ument

0.98

eries

0.89

ito

0.84

ingly

0.83

atory

0.82

ery

0.81

eting

0.79

 mock

0.77

tails

0.75

Activations Density 0.029%

references to simulated or imitation versions of something

references to "mockumentaries" or similar formats

No Comments

No Known Activations