INDEX

Explanations

doubts or reservations expressed about potential actions or statements

oai_token-act-pair · gpt-3.5-turbo

phrasings emphasizing the phrase "not that."

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ãĥ«

-0.69

Ö¼

-0.68

orah

-0.67

 Souls

-0.63

 Peaks

-0.63

oS

-0.63

mates

-0.63

Spawn

-0.62

ãĥĳ

-0.60

bage

-0.59

POSITIVE LOGITS

 pesky

0.92

cher

0.83

 bothered

0.80

 mattered

0.77

 fateful

0.77

ndra

0.77

interstitial

0.76

angular

0.75

 bothers

0.75

 suspic

0.75

Activations Density 0.178%

doubts or reservations expressed about potential actions or statements

phrasings emphasizing the phrase "not that."

No Comments

No Known Activations