INDEX

Explanations

mention of the ability to perform actions or tasks

oai_token-act-pair · gpt-3.5-turbo

instances of the word "can."

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

edient

-0.65

 Fighter

-0.63

 rehearsal

-0.63

 Mant

-0.60

Ing

-0.58

ependence

-0.57

gement

-0.56

ciplinary

-0.55

 guarding

-0.55

Moz

-0.54

POSITIVE LOGITS

't

1.35

adian

1.16

berra

1.02

NOT

0.94

tera

0.92

vas

0.90

 attest

0.88

ny

0.82

thus

0.75

nery

0.74

Activations Density 0.128%

mention of the ability to perform actions or tasks

instances of the word "can."

No Comments

No Known Activations