INDEX

Explanations

numbers and numerical values

oai_token-act-pair · gpt-3.5-turbo

numerical data and estimates typically associated with statistics or predictions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 beauty

-0.68

 bombard

-0.65

 roses

-0.62

 liberating

-0.62

kit

-0.62

 swim

-0.61

 beaut

-0.60

 stre

-0.60

 bunny

-0.59

 dressing

-0.59

POSITIVE LOGITS

1.32

1.23

1.19

1.16

1.15

1.14

1.11

Activations Density 0.094%

numbers and numerical values

numerical data and estimates typically associated with statistics or predictions

No Comments

No Known Activations