INDEX

Explanations

words like "admit", "confess", and "say" indicating acknowledgment or disclosure

oai_token-act-pair · gpt-3.5-turbo

expressions of acknowledgment or admission of personal feelings or actions

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 decom

-0.71

locked

-0.66

ishable

-0.64

let

-0.61

stall

-0.61

 lockdown

-0.59

lets

-0.57

arent

-0.57

 scrimmage

-0.57

 resur

-0.57

POSITIVE LOGITS

 anecd

0.77

 hindsight

0.76

eno

0.74

fav

0.68

 admit

0.68

vv

0.63

lers

0.62

olson

0.61

erning

0.61

Uri

0.61

Activations Density 0.052%

words like "admit", "confess", and "say" indicating acknowledgment or disclosure

expressions of acknowledgment or admission of personal feelings or actions

No Comments

No Known Activations