INDEX

Explanations

mentions of the names of individuals, particularly with a focus on medical professionals and politicians

oai_token-act-pair · gpt-3.5-turbo

variations of the word "medium" and its related forms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ĸļ

-0.99

 Ember

-0.70

``

-0.68

 Crom

-0.62

©¶æ

-0.61

 Chrys

-0.61

omorph

-0.60

 disbelief

-0.60

semble

-0.60

 Jasper

-0.60

POSITIVE LOGITS

ocre

0.84

illions

0.81

geist

0.75

erver

0.65

aband

0.65

icum

0.62

owler

0.62

get

0.61

 tending

0.61

ocr

0.61

Activations Density 0.093%

mentions of the names of individuals, particularly with a focus on medical professionals and politicians

variations of the word "medium" and its related forms

No Comments

No Known Activations