INDEX

Explanations

references to various types of disabilities

oai_token-act-pair · gpt-3.5-turbo

references to learning disabilities and related conditions

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

RFC

-0.73

-+-+-+-+

-0.69

 ellipt

-0.69

 Camel

-0.64

ICAN

-0.62

Population

-0.62

CBI

-0.61

Moscow

-0.60

Sed

-0.59

 Grove

-0.59

POSITIVE LOGITS

 disabilities

1.23

 impair

1.06

hops

0.91

 inval

0.86

ometimes

0.85

hips

0.82

 traged

0.82

quished

0.80

 impaired

0.80

anie

0.78

Activations Density 0.018%

references to various types of disabilities

references to learning disabilities and related conditions

No Comments

No Known Activations