INDEX

Explanations

people's names or words involving names that contain "ha."

oai_token-act-pair · gpt-3.5-turbo

instances of the term "ha" in various contexts, likely indicating laughter or amusement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

atories

-0.84

 outgoing

-0.70

hips

-0.66

brain

-0.63

tailed

-0.62

atory

-0.61

awaru

-0.60

inarily

-0.59

 outfield

-0.58

 treating

-0.58

POSITIVE LOGITS

wn

1.33

user

1.14

illard

1.00

irst

0.95

azard

0.94

emi

0.89

ifa

0.87

ilee

0.86

pless

0.85

 Pradesh

0.85

Activations Density 0.033%

people's names or words involving names that contain "ha."

instances of the term "ha" in various contexts, likely indicating laughter or amusement

No Comments

No Known Activations