INDEX

Explanations

terms related to foreign languages or potentially a mix of different characters/symbols

oai_token-act-pair · gpt-3.5-turbo

special characters or symbols that may indicate formatting or encoding issues

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

ividual

-0.77

hower

-0.75

earcher

-0.74

elsen

-0.73

krit

-0.69

urers

-0.68

urally

-0.67

orial

-0.66

onomy

-0.65

ettings

-0.63

POSITIVE LOGITS

¯

1.06

Ĩ

0.93

İ

0.92

à¤¾

0.90

¦

0.87

®

0.86

į

0.86

¾

0.84

°

0.83

º

0.81

Activations Density 0.006%

terms related to foreign languages or potentially a mix of different characters/symbols

special characters or symbols that may indicate formatting or encoding issues

No Comments

No Known Activations