INDEX

Explanations

facts or information

oai_token-act-pair · gpt-3.5-turbo

assertions of knowledge or certainty

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

pex

-0.69

cit

-0.68

nesota

-0.67

isco

-0.64

cific

-0.64

 offic

-0.64

coni

-0.64

conservancy

-0.63

ksh

-0.63

oshenko

-0.63

POSITIVE LOGITS

ledged

0.95

ledge

0.92

edge

0.77

nothing

0.76

 ourselves

0.75

lege

0.75

how

0.71

LED

0.70

 beforehand

0.68

 nothing

0.65

Activations Density 0.048%

facts or information

assertions of knowledge or certainty

No Comments

No Known Activations