INDEX

Explanations

phrases related to evaluating the effectiveness or consequences of actions

oai_token-act-pair · gpt-3.5-turbo

themes of uncertainty and evaluation of evidence or policies

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 reminds

-0.60

inis

-0.58

uploads

-0.58

anwhile

-0.57

razil

-0.57

lda

-0.55

 Valentine

-0.53

 amazed

-0.51

 reminded

-0.51

azon

-0.50

POSITIVE LOGITS

 whatsoever

0.70

 worthwhile

0.66

CHQ

0.63

 desired

0.62

 outcome

0.61

izoph

0.60

)).

0.59

 meaningful

0.59

 harm

0.58

ratom

0.58

Activations Density 1.160%

phrases related to evaluating the effectiveness or consequences of actions

themes of uncertainty and evaluation of evidence or policies

No Comments

No Known Activations