INDEX

Explanations

names or titles within text

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "name" and its context within a discussion

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

gif

-0.70

icult

-0.69

isexual

-0.66

EMS

-0.63

ureau

-0.62

 Dragonbound

-0.62

athing

-0.62

aples

-0.62

thumbnails

-0.62

erker

-0.60

POSITIVE LOGITS

plate

1.34

plates

1.25

 recognition

0.90

paces

0.89

tag

0.78

paced

0.76

 synonymous

0.75

calling

0.74

names

0.73

ames

0.73

Activations Density 0.049%

names or titles within text

occurrences of the word "name" and its context within a discussion

No Comments

No Known Activations