INDEX

Explanations

anecdotal evidence in support of various claims

oai_token-act-pair · gpt-3.5-turbo

references to evidence and research studies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

||||

-0.77

 ç¥ŀ

-0.76

ategor

-0.70

ndra

-0.69

bent

-0.68

yles

-0.67

rex

-0.66

paces

-0.65

[|

-0.65

reddits

-0.64

POSITIVE LOGITS

 evidence

1.97

evidence

1.84

 indications

1.57

Evidence

1.55

 Evidence

1.47

 evid

1.43

 testimony

1.41

 anecdotal

1.40

 testimonies

1.36

 reports

1.33

Activations Density 1.123%

anecdotal evidence in support of various claims

references to evidence and research studies

No Comments

No Known Activations