INDEX

Explanations

numerical totals or sums

oai_token-act-pair · gpt-3.5-turbo

numerical totals and aggregates

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

afe

-0.69

 Portug

-0.66

irk

-0.64

aptic

-0.63

elligent

-0.62

 Bunker

-0.59

 tongue

-0.59

guided

-0.58

Pix

-0.58

advertisement

-0.57

POSITIVE LOGITS

 totals

3.54

 totaled

2.57

 averages

2.13

 totaling

2.13

 total

1.61

total

1.55

Total

1.47

 tallied

1.44

 tally

1.39

 Total

1.36

Activations Density 0.020%

numerical totals or sums

numerical totals and aggregates

No Comments

No Known Activations