INDEX

Explanations

dates expressed in ordinal form (ending in 'th', 'st', 'nd', 'rd')

oai_token-act-pair · gpt-3.5-turbo

occurrences of dates or ordinal indicators

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

Marginal

-0.69

opic

-0.58

 pedigree

-0.57

 temperament

-0.57

 tame

-0.57

mbuds

-0.56

ufact

-0.55

 indebted

-0.55

 interference

-0.55

 fertile

-0.54

POSITIVE LOGITS

1.02

0.96

0.89

0.87

ruary

0.83

0.81

 Annual

0.80

Activations Density 0.022%

dates expressed in ordinal form (ending in 'th', 'st', 'nd', 'rd')

occurrences of dates or ordinal indicators

No Comments

No Known Activations