INDEX

Explanations

mentions of abbreviations or acronyms

oai_token-act-pair · gpt-3.5-turbo

instances of the abbreviation "ED" followed by numbers or related content

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

ãĥ£

-0.74

fruit

-0.68

include

-0.68

ãĤ§

-0.66

 Chak

-0.64

ãĤ©

-0.64

juries

-0.62

vim

-0.61

jured

-0.60

aird

-0.60

POSITIVE LOGITS

dit

1.10

ERAL

1.00

uce

1.00

nesday

0.98

RIC

0.98

IUM

0.96

GE

0.94

aily

0.94

ouble

0.93

IAS

0.92

Activations Density 0.014%

mentions of abbreviations or acronyms

instances of the abbreviation "ED" followed by numbers or related content

No Comments

No Known Activations