INDEX

Explanations

references to fires or burning

oai_token-act-pair · gpt-3.5-turbo

references to the word "burning" in various contexts

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

ois

-0.89

berman

-0.74

 Inform

-0.72

clusive

-0.72

 Greenberg

-0.71

roup

-0.71

 Pros

-0.68

 Pearce

-0.68

brates

-0.67

ante

-0.65

POSITIVE LOGITS

 burning

3.76

burning

2.73

 burn

2.36

 burned

2.31

 burns

2.04

 Burning

2.04

 burnt

2.00

Burn

1.87

 burner

1.70

 flaming

1.60

Activations Density 0.013%

references to fires or burning

references to the word "burning" in various contexts

No Comments

No Known Activations