INDEX

Explanations

instances where praise or recognition is given

oai_token-act-pair · gpt-3.5-turbo

expressions of acknowledgment or praise

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

gren

-0.66

 fronts

-0.65

tered

-0.64

few

-0.63

 lyrics

-0.63

 rall

-0.63

ptic

-0.62

 WRITE

-0.61

 Beng

-0.61

glers

-0.60

POSITIVE LOGITS

udos

1.13

elta

0.92

ĺħ

0.91

itto

0.90

agram

0.87

obos

0.86

insula

0.85

ItemTracker

0.85

utra

0.83

wcsstore

0.82

Activations Density 0.038%

instances where praise or recognition is given

expressions of acknowledgment or praise

No Comments

No Known Activations