INDEX

Explanations

adjectives ending in '-ful' or '-ing'

oai_token-act-pair · gpt-3.5-turbo

words that signify various forms of suffixes typically indicating attributes or qualities

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

 Liberties

-0.68

 Dragonbound

-0.64

 DATA

-0.63

 Ashes

-0.59

 whine

-0.57

 equivalents

-0.57

 surrogate

-0.56

 Standards

-0.56

 ceased

-0.55

enhagen

-0.55

POSITIVE LOGITS

rill

0.96

opter

0.92

gment

0.91

eks

0.87

olon

0.85

phal

0.84

omi

0.84

ules

0.83

¬

0.82

»

0.82

Activations Density 0.299%

adjectives ending in '-ful' or '-ing'

words that signify various forms of suffixes typically indicating attributes or qualities

No Comments

No Known Activations