INDEX

Explanations

phrases indicating observations or examinations

oai_token-act-pair · gpt-3.5-turbo

personal pronouns and references to viewers or audiences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Virtue

-0.69

 Vengeance

-0.63

 delaying

-0.62

 Kills

-0.62

 extortion

-0.61

 sacrific

-0.61

 ransom

-0.59

 publicity

-0.59

 onstage

-0.58

yss

-0.58

POSITIVE LOGITS

uggest

0.99

 discern

0.91

 understand

0.85

 realise

0.84

 noticed

0.82

inguishable

0.81

 realize

0.79

 glimpse

0.79

RESULTS

0.79

aeus

0.77

Activations Density 0.348%

phrases indicating observations or examinations

personal pronouns and references to viewers or audiences

No Comments

No Known Activations