INDEX

Explanations

adverbs denoting intensity

oai_token-act-pair · gpt-3.5-turbo

intense qualifiers indicating strong emphasis or importance

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

peak

-0.69

amel

-0.64

isec

-0.63

Afee

-0.60

bringing

-0.59

layer

-0.58

loads

-0.57

gotten

-0.56

ãĥ¼ãĥĨãĤ£

-0.56

runs

-0.56

POSITIVE LOGITS

 much

1.02

 seldom

0.94

 rarely

0.86

 important

0.85

 similar

0.84

 different

0.83

 difficult

0.82

 unlikely

0.82

 likely

0.79

much

0.78

Activations Density 0.088%

adverbs denoting intensity

intense qualifiers indicating strong emphasis or importance

No Comments

No Known Activations