INDEX

Explanations

phrases related to being rewarded or greeted with something

oai_token-act-pair · gpt-3.5-turbo

phrases indicating a response or result related to receiving something

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

wake

-0.76

sche

-0.68

urat

-0.68

wash

-0.66

soon

-0.64

 deceived

-0.62

alias

-0.61

 susceptible

-0.59

affected

-0.58

 decom

-0.58

POSITIVE LOGITS

Interstitial

0.78

arde

0.74

Ĥª

0.72

ļéĨĴ

0.69

 applause

0.69

eering

0.69

arrass

0.69

 cheers

0.69

¶æ

0.69

æµ

0.68

Activations Density 0.312%

phrases related to being rewarded or greeted with something

phrases indicating a response or result related to receiving something

No Comments

No Known Activations