INDEX

Explanations

proper nouns or specific terms related to companies or organizations

oai_token-act-pair · gpt-3.5-turbo

references to corporate entities

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

eful

-0.86

ppings

-0.82

req

-0.81

efully

-0.75

joy

-0.74

imm

-0.73

xual

-0.73

uten

-0.73

rose

-0.73

lyn

-0.73

POSITIVE LOGITS

wide

1.04

ridor

0.90

orporated

0.86

orpor

0.85

 Headquarters

0.83

 Counsel

0.82

 headquartered

0.79

LTD

0.79

oreal

0.78

eers

0.77

Activations Density 0.012%

proper nouns or specific terms related to companies or organizations

references to corporate entities

No Comments

No Known Activations