INDEX

Explanations

numbers or quantities mentioned in the text

oai_token-act-pair · gpt-3.5-turbo

quantitative descriptions or approximations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 Peaks

-0.77

oly

-0.70

Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯

-0.69

ãĤ¿

-0.69

CHR

-0.67

rift

-0.67

SS

-0.67

à¥

-0.66

SPONSORED

-0.66

Write

-0.65

POSITIVE LOGITS

 halfway

0.84

 midway

0.69

PsyNetMessage

0.68

0.67

aleb

0.66

 calib

0.65

 migrating

0.65

 blowing

0.64

 ducks

0.64

 reforming

0.64

Activations Density 0.156%

numbers or quantities mentioned in the text

quantitative descriptions or approximations

No Comments

No Known Activations