INDEX

Explanations

phrases describing newly acquired knowledge or information

oai_token-act-pair · gpt-3.5-turbo

expressions of acquired knowledge or information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

abwe

-0.74

oided

-0.71

 throats

-0.65

 pers

-0.64

idity

-0.64

adies

-0.62

ankind

-0.62

ataka

-0.61

yip

-0.58

vell

-0.58

POSITIVE LOGITS

llor

0.90

Lear

0.81

ilage

0.79

lyn

0.77

çīĪ

0.75

æ©

0.75

 firsthand

0.75

æĥ

0.73

CLASSIFIED

0.73

learn

0.72

Activations Density 0.028%

phrases describing newly acquired knowledge or information

expressions of acquired knowledge or information

No Comments

No Known Activations