INDEX

Explanations

proper nouns or titles with a capital letter followed by various characters

oai_token-act-pair · gpt-3.5-turbo

references to viewing, perspectives, and surveillance-related topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

shr

-0.84

pmwiki

-0.79

 taxing

-0.77

 vain

-0.76

secut

-0.76

 stocking

-0.75

 restraining

-0.74

 muff

-0.72

 mysterious

-0.69

 sporadic

-0.69

POSITIVE LOGITS

iew

1.70

Sov

0.96

icz

0.89

allery

0.89

ert

0.86

ski

0.83

atio

0.83

erm

0.82

escription

0.81

sburgh

0.81

Activations Density 0.008%

proper nouns or titles with a capital letter followed by various characters

references to viewing, perspectives, and surveillance-related topics

No Comments

No Known Activations