INDEX

Explanations

winning or losing streaks in sports or competitive contexts

oai_token-act-pair · gpt-3.5-turbo

mentions of winning or losing streaks in performance

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

ãĤº

-0.79

abad

-0.72

resources

-0.70

Male

-0.69

ACTED

-0.68

Lod

-0.66

gency

-0.66

idd

-0.65

poses

-0.64

onomic

-0.64

POSITIVE LOGITS

 streak

1.23

iness

0.88

manship

0.87

uous

0.84

icity

0.82

 streaks

0.81

isSpecialOrderable

0.76

wcs

0.69

yk

0.69

terness

0.68

Activations Density 0.026%

winning or losing streaks in sports or competitive contexts

mentions of winning or losing streaks in performance

No Comments

No Known Activations