INDEX

Explanations

phrases related to achieving victory or success

oai_token-act-pair · gpt-3.5-turbo

instances of the word "win."

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

alam

-0.70

erity

-0.68

rouch

-0.64

urches

-0.63

umn

-0.63

udes

-0.62

Uz

-0.61

agg

-0.59

 condu

-0.58

 Dwell

-0.58

POSITIVE LOGITS

throp

0.81

now

0.78

ners

0.73

iem

0.71

nings

0.70

riors

0.69

atown

0.69

stroke

0.68

ced

0.68

 Citation

0.66

Activations Density 0.038%

phrases related to achieving victory or success

instances of the word "win."

No Comments

No Known Activations