INDEX

Explanations

words that are abbreviated within sentences using "i.e."

oai_token-act-pair · gpt-3.5-turbo

the letter "e" in various contexts throughout the document

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

 crunch

-0.65

rament

-0.64

 mids

-0.60

Round

-0.60

 neutrality

-0.58

rush

-0.57

Crunch

-0.56

 seams

-0.56

ingham

-0.56

 Flavoring

-0.56

POSITIVE LOGITS

clair

0.98

.,

0.87

clusive

0.83

lements

0.82

gypt

0.79

urance

0.79

lev

0.74

lege

0.74

lement

0.72

viously

0.72

Activations Density 0.014%

words that are abbreviated within sentences using "i.e."

the letter "e" in various contexts throughout the document

No Comments

No Known Activations