INDEX

Explanations

phrases indicating perception or judgment, particularly using the word "impression."

oai_token-act-pair · gpt-3.5-turbo

phrases that convey perceived impressions or opportunities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

adow

-0.63

 Speedway

-0.60

ICO

-0.59

 pursuits

-0.56

 forthcoming

-0.54

ilver

-0.54

arer

-0.54

verning

-0.53

held

-0.52

 practition

-0.52

POSITIVE LOGITS

elight

0.76

liest

0.75

ultimate

0.74

iest

0.70

 anew

0.69

same

0.65

forts

0.65

UTH

0.63

erity

0.61

Reloaded

0.60

Activations Density 0.289%

phrases indicating perception or judgment, particularly using the word "impression."

phrases that convey perceived impressions or opportunities

No Comments

No Known Activations