INDEX

Explanations

verbs or phrases indicating a distinction or uniqueness

oai_token-act-pair · gpt-3.5-turbo

questions about differentiation or what makes something unique

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

abad

-0.85

bill

-0.79

lance

-0.77

lex

-0.75

lla

-0.72

ixel

-0.69

estern

-0.68

endix

-0.67

ammy

-0.66

jab

-0.66

POSITIVE LOGITS

 motiv

0.80

pires

0.78

 motivating

0.75

 Unleashed

0.70

 distinguishes

0.70

rament

0.69

 motivate

0.68

 GOODMAN

0.66

piring

0.66

SPONSORED

0.64

Activations Density 0.098%

verbs or phrases indicating a distinction or uniqueness

questions about differentiation or what makes something unique

No Comments

No Known Activations