INDEX

Explanations

references to the word "ca."

oai_token-act-pair · gpt-3.5-turbo

references to aviation incidents or tragedies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

schild

-0.79

states

-0.78

pocket

-0.78

 GOODMAN

-0.74

mble

-0.72

tk

-0.69

lings

-0.66

hander

-0.65

UID

-0.64

rats

-0.64

POSITIVE LOGITS

ption

1.23

esar

1.09

utical

1.00

ffe

0.87

ibo

0.87

UTION

0.85

ution

0.84

usa

0.83

iba

0.83

zza

0.81

Activations Density 0.012%

references to the word "ca."

references to aviation incidents or tragedies

No Comments

No Known Activations