INDEX

Explanations

references to fires

oai_token-act-pair · gpt-3.5-turbo

mentions of fire-related incidents or terms

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

xus

-0.83

 Virtue

-0.80

 Vide

-0.80

 Freed

-0.75

Tok

-0.71

laus

-0.70

amaru

-0.69

atem

-0.68

afort

-0.67

Birth

-0.65

POSITIVE LOGITS

flies

1.13

storm

1.06

 exting

1.05

fighting

1.00

proof

0.96

storms

0.95

fight

0.93

trap

0.92

bomb

0.92

balls

0.90

Activations Density 0.022%

references to fires

mentions of fire-related incidents or terms

No Comments

No Known Activations