INDEX

Explanations

big or large numerical values or quantities

oai_token-act-pair · gpt-3.5-turbo

words that indicate size or magnitude, specifically emphasizing "big" and "large."

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

casts

-0.87

iddles

-0.86

agents

-0.86

codes

-0.85

brids

-0.84

iths

-0.83

speak

-0.82

rants

-0.79

laws

-0.78

ults

-0.78

POSITIVE LOGITS

 swath

1.07

 chunk

1.06

 amount

0.98

 dose

0.89

 headache

0.87

 portion

0.86

 subset

0.86

 number

0.85

 degree

0.83

 uptick

0.81

Activations Density 0.153%

big or large numerical values or quantities

words that indicate size or magnitude, specifically emphasizing "big" and "large."

No Comments

No Known Activations