INDEX

Explanations

references to jokes or humorous statements

oai_token-act-pair · gpt-3.5-turbo

mentions of jokes or humorous references

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

phalt

-0.83

undai

-0.82

porting

-0.72

enture

-0.72

CLASSIFIED

-0.69

rive

-0.68

ignty

-0.66

ighting

-0.66

orneys

-0.66

ports

-0.66

POSITIVE LOGITS

ously

0.99

 jokes

0.91

 joking

0.86

 joke

0.86

bags

0.81

bag

0.79

osal

0.79

 caller

0.77

Pom

0.77

 mocking

0.76

Activations Density 0.019%

references to jokes or humorous statements

mentions of jokes or humorous references

No Comments

No Known Activations