INDEX

Explanations

sequences of characters in a foreign language, possibly indicating text encoding or corruption

oai_token-act-pair · gpt-3.5-turbo

symbols or special characters used in text

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

ierrez

-0.85

ministic

-0.72

aminer

-0.71

 ethic

-0.71

achus

-0.69

ebus

-0.69

ilater

-0.67

lopp

-0.66

ettings

-0.65

ription

-0.63

POSITIVE LOGITS

¾

1.16

ħ

1.09

Ĩ

0.98

¼

0.96

Û

0.96

µ

0.94

°

0.94

¹

0.93

Ĥª

0.92

Į

0.90

Activations Density 0.003%

sequences of characters in a foreign language, possibly indicating text encoding or corruption

symbols or special characters used in text

No Comments

No Known Activations