INDEX

Explanations

phrases related to categories or groups of people or things

oai_token-act-pair · gpt-3.5-turbo

terms related to collective groups and responsibilities

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

edia

-0.72

ahime

-0.70

agger

-0.65

epad

-0.65

tch

-0.63

instein

-0.63

 Braun

-0.63

ews

-0.62

 proverb

-0.61

uristic

-0.60

POSITIVE LOGITS

 except

1.53

except

1.20

 imaginable

1.07

 whatsoever

1.03

 alike

1.01

 irrespective

0.96

soever

0.94

 facets

0.89

 thereafter

0.81

 regardless

0.80

Activations Density 0.308%

phrases related to categories or groups of people or things

terms related to collective groups and responsibilities

No Comments

No Known Activations