INDEX

Explanations

phrases related to being selective or having guidelines

oai_token-act-pair · gpt-3.5-turbo

references to criteria and guidelines for decision-making

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

irit

-0.83

lez

-0.81

vind

-0.81

fort

-0.79

olitics

-0.79

loo

-0.77

USS

-0.75

Champ

-0.74

KC

-0.74

anoia

-0.72

POSITIVE LOGITS

how

1.30

 assigning

1.30

 selecting

1.29

 specifying

1.23

 determining

1.20

 awarding

1.19

 choosing

1.18

 disclosing

1.18

 naming

1.13

 accessing

1.11

Activations Density 0.317%

phrases related to being selective or having guidelines

references to criteria and guidelines for decision-making

No Comments

No Known Activations