INDEX

Explanations

instances where something is especially vulnerable or susceptible to certain conditions or events

oai_token-act-pair · gpt-3.5-turbo

the word "prone" in various contexts related to risk or vulnerability

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ACA

-0.79

tein

-0.78

 CLSID

-0.75

ILE

-0.73

ZA

-0.70

âĸĳâĸĳ

-0.70

zyme

-0.67

GROUP

-0.67

elle

-0.66

OTOS

-0.65

POSITIVE LOGITS

 prone

1.18

entimes

1.04

 challeng

1.02

 confir

1.01

 tremend

0.96

 condem

0.90

 experien

0.86

 mathemat

0.86

 conduc

0.85

 compr

0.85

Activations Density 0.008%

instances where something is especially vulnerable or susceptible to certain conditions or events

the word "prone" in various contexts related to risk or vulnerability

No Comments

No Known Activations