INDEX

Explanations

[action verbs expressed in the past tense related to various activities or events.]

oai_token-act-pair · gpt-3.5-turbo

instances of the word "took."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Smile

-0.66

lex

-0.66

Cong

-0.66

ler

-0.64

 agre

-0.60

 icing

-0.59

bie

-0.58

 ---------

-0.58

psey

-0.58

gom

-0.58

POSITIVE LOGITS

 advantage

1.15

aways

1.11

 aback

1.02

 refuge

1.00

 pains

0.95

 precedence

0.92

 place

0.89

 care

0.86

 heed

0.85

prising

0.84

Activations Density 0.090%

[action verbs expressed in the past tense related to various activities or events.]

instances of the word "took."

No Comments

No Known Activations