INDEX

Explanations

keywords related to burns and burning in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to fire and burning

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

FN

-0.80

reme

-0.73

omm

-0.72

 Zimmer

-0.71

 Introduced

-0.69

artisan

-0.69

 introduced

-0.68

 Marketplace

-0.67

sup

-0.65

anto

-0.64

POSITIVE LOGITS

 burn

3.81

 burns

2.82

 burned

2.47

 burning

2.33

Burn

2.26

 Burn

2.22

burn

2.13

burning

2.08

 burner

2.06

 burnt

1.98

Activations Density 0.009%

keywords related to burns and burning in various contexts

references to fire and burning

No Comments

No Known Activations