INDEX

Explanations

quotations within text, specifically attributions indicated by the word "said."

oai_token-act-pair · gpt-3.5-turbo

instances of dialogue or direct speech

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

phia

-0.63

Appearances

-0.62

ICES

-0.61

Pont

-0.58

uncture

-0.56

 wearer

-0.56

 Barbarian

-0.55

 convergence

-0.55

 Skydragon

-0.54

HP

-0.54

POSITIVE LOGITS

 sarcast

1.24

 bluntly

1.01

 aloud

1.01

 rhet

0.94

 goodbye

0.94

 confidently

0.87

 emphatically

0.87

 proudly

0.86

 bitterly

0.86

 politely

0.81

Activations Density 0.152%

quotations within text, specifically attributions indicated by the word "said."

instances of dialogue or direct speech

No Comments

No Known Activations