INDEX

Explanations

mentions of specific individuals by their last name or full name

oai_token-act-pair · gpt-3.5-turbo

references to individuals and their associated actions or statements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

Measure

-0.62

 Incre

-0.62

 transformative

-0.61

 Dram

-0.61

 Calcul

-0.60

 trillions

-0.60

 enactment

-0.59

 topple

-0.58

 endemic

-0.58

 analytical

-0.57

POSITIVE LOGITS

 declined

1.21

 hasn

1.19

 insists

1.18

 denies

1.17

 tweeted

1.12

 reportedly

1.11

 assures

1.10

 intends

1.09

 spokesperson

1.08

 remains

1.07

Activations Density 0.427%

mentions of specific individuals by their last name or full name

references to individuals and their associated actions or statements

No Comments

No Known Activations