INDEX

Explanations

words related to physical impacts or collisions

oai_token-act-pair · gpt-3.5-turbo

references to physical contact or collisions

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

natureconservancy

-0.74

 Ecology

-0.72

 Tigers

-0.66

xious

-0.65

ANCE

-0.65

eval

-0.63

encia

-0.62

ndra

-0.62

LAW

-0.61

umerable

-0.61

POSITIVE LOGITS

 bump

1.03

olean

0.95

 bumps

0.92

downs

0.87

balls

0.82

alon

0.80

engers

0.76

alos

0.76

sticks

0.75

ing

0.73

Activations Density 0.010%

words related to physical impacts or collisions

references to physical contact or collisions

No Comments

No Known Activations