INDEX

Explanations

mentions of returning to a particular place or position

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "to."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 conveyed

-0.73

croft

-0.72

SourceFile

-0.69

ially

-0.67

 gone

-0.66

 communicated

-0.65

achy

-0.64

 Advertisement

-0.61

erity

-0.61

*/(

-0.60

POSITIVE LOGITS

 haunt

1.11

 shore

0.89

 fold

0.81

 pasture

0.81

 prominence

0.77

lict

0.72

isconsin

0.72

 fray

0.70

 earth

0.70

 Valhalla

0.69

Activations Density 0.056%

mentions of returning to a particular place or position

occurrences of the word "to."

No Comments

No Known Activations