INDEX

Explanations

mentions of specific details or facts within a broader context

oai_token-act-pair · gpt-3.5-turbo

instances of the phrase "noting that."

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

 Escape

-0.56

 live

-0.56

angle

-0.55

als

-0.55

ular

-0.54

 lives

-0.54

 Hack

-0.54

oi

-0.53

cles

-0.52

ug

-0.52

POSITIVE LOGITS

 noting

3.21

 mentioning

2.04

 citing

1.86

 stating

1.84

 emphasizing

1.83

 acknowledging

1.83

 stressing

1.72

 pointing

1.71

 noticing

1.67

 estimating

1.58

Activations Density 0.009%

mentions of specific details or facts within a broader context

instances of the phrase "noting that."

No Comments

No Known Activations