INDEX

Explanations

terms related to negative connotations or derogatory expressions

oai_token-act-pair · gpt-3.5-turbo

variations of the word "moron."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

BOOK

-0.80

ACTED

-0.67

orney

-0.64

payer

-0.64

XY

-0.63

 Wheeler

-0.62

 Hawaiian

-0.62

ECH

-0.61

sets

-0.60

Seattle

-0.60

POSITIVE LOGITS

gue

1.06

aceutical

0.99

gage

0.92

mor

0.91

iday

0.86

ose

0.82

idian

0.81

icultural

0.80

itary

0.79

phe

0.79

Activations Density 0.008%

terms related to negative connotations or derogatory expressions

variations of the word "moron."

No Comments

No Known Activations