INDEX

Explanations

terms related to division or disagreement among groups or individuals

oai_token-act-pair · gpt-3.5-turbo

instances of the word "divided" in various contexts

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

ORTS

-0.72

queue

-0.67

 Fargo

-0.66

ãĥīãĥ©

-0.66

ECH

-0.65

OD

-0.64

EEK

-0.64

rouse

-0.64

CHA

-0.64

eno

-0.63

POSITIVE LOGITS

 divided

0.80

 evenly

0.75

rescent

0.75

SPL

0.72

naire

0.71

iple

0.70

reth

0.70

rified

0.69

 unequ

0.67

ocument

0.67

Activations Density 0.020%

terms related to division or disagreement among groups or individuals

instances of the word "divided" in various contexts

No Comments

No Known Activations