INDEX

Explanations

phrases with pronouns indicating specific individuals

oai_token-act-pair · gpt-3.5-turbo

pronouns, particularly those referring to male individuals

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

 Fail

-0.65

fail

-0.59

 Killing

-0.58

 Underworld

-0.58

 Anarchy

-0.57

Row

-0.55

 Description

-0.55

Chem

-0.55

metal

-0.55

 Destruction

-0.55

POSITIVE LOGITS

 expects

1.39

 understands

1.29

 believes

1.28

 intends

1.25

 regretted

1.21

 thinks

1.20

 disagrees

1.20

 hoped

1.19

 hopes

1.18

 regrets

1.18

Activations Density 0.125%

phrases with pronouns indicating specific individuals

pronouns, particularly those referring to male individuals

No Comments

No Known Activations