INDEX

Explanations

statements where someone is speaking or reporting information

oai_token-act-pair · gpt-3.5-turbo

instances of speech attribution, particularly the word "says."

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

cffffcc

-0.67

nerg

-0.64

 blows

-0.63

inary

-0.63

aired

-0.62

Holy

-0.62

Unity

-0.62

Chan

-0.60

respective

-0.60

ugal

-0.60

POSITIVE LOGITS

omething

0.90

ometimes

0.86

olate

0.83

paces

0.83

ynthesis

0.80

creen

0.79

hiba

0.76

ysis

0.74

 goodbye

0.74

earchers

0.72

Activations Density 0.134%

statements where someone is speaking or reporting information

instances of speech attribution, particularly the word "says."

No Comments

No Known Activations