INDEX

Explanations

abbreviations and acronyms related to technical and medical terms

oai_token-act-pair · gpt-3.5-turbo

references to the abbreviation "MT" or similar sequences in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

hold

-0.86

holes

-0.68

piring

-0.64

cific

-0.63

pta

-0.63

 cider

-0.63

Boe

-0.63

 Spencer

-0.62

alyses

-0.61

pects

-0.61

POSITIVE LOGITS

MT

1.01

ypes

0.82

ument

0.78

TON

0.77

TL

0.76

GO

0.75

orsi

0.75

ORPG

0.73

BF

0.73

ullah

0.73

Activations Density 0.009%

abbreviations and acronyms related to technical and medical terms

references to the abbreviation "MT" or similar sequences in various contexts

No Comments

No Known Activations