INDEX

Explanations

weight measurements, often in kilograms

oai_token-act-pair · gpt-3.5-turbo

terms related to imaging or photographic technology

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 perks

-0.65

ACC

-0.64

 breaches

-0.63

irk

-0.61

psc

-0.61

Cav

-0.59

 cant

-0.58

 Added

-0.58

 Channel

-0.58

Adds

-0.58

POSITIVE LOGITS

ogram

4.61

ograms

3.75

ography

1.65

ographic

1.65

ograph

1.65

ometry

1.50

rogram

1.50

ographs

1.50

ographically

1.46

ograp

1.36

Activations Density 0.017%

weight measurements, often in kilograms

terms related to imaging or photographic technology

No Comments

No Known Activations