INDEX

Explanations

specific named entities like "documents."

oai_token-act-pair · gpt-3.5-turbo

references to official records or files

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

twitch

-0.74

reciation

-0.71

pit

-0.66

ensity

-0.66

rossover

-0.65

obic

-0.65

alty

-0.65

=#

-0.64

chance

-0.63

olation

-0.63

POSITIVE LOGITS

 documents

3.62

 Documents

2.69

 document

2.42

Documents

2.33

 papers

2.22

 documentation

1.86

 docs

1.84

 memos

1.84

 records

1.80

Document

1.73

Activations Density 0.016%

specific named entities like "documents."

references to official records or files

No Comments

No Known Activations