INDEX

Explanations

informal language involving places or organizations, like nicknames or abbreviations

oai_token-act-pair · gpt-3.5-turbo

places and organizations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

spring

-0.80

ERN

-0.76

ELL

-0.72

andise

-0.71

orage

-0.69

ebted

-0.69

ifies

-0.69

eers

-0.68

izations

-0.68

ifying

-0.67

POSITIVE LOGITS

volent

0.82

vious

0.80

judicial

0.77

gress

0.76

phrine

0.76

AFTA

0.72

llor

0.69

mand

0.69

pless

0.69

boro

0.67

Activations Density 0.027%

informal language involving places or organizations, like nicknames or abbreviations

places and organizations

No Comments

No Known Activations