INDEX

Explanations

phrases or excerpts that emphasize precision or accuracy

oai_token-act-pair · gpt-3.5-turbo

phrases emphasizing precision or accuracy

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Flavoring

-0.78

stown

-0.69

ker

-0.69

kj

-0.68

 Democr

-0.68

 Krishna

-0.68

GD

-0.68

 Psychiat

-0.67

 Patricia

-0.66

Reviewer

-0.65

POSITIVE LOGITS

 exact

1.12

itude

0.93

itudes

0.78

 wording

0.78

 amount

0.78

entimes

0.75

 ident

0.73

edIn

0.73

orius

0.72

aneously

0.72

Activations Density 0.005%

phrases or excerpts that emphasize precision or accuracy

phrases emphasizing precision or accuracy

No Comments

No Known Activations