INDEX

Explanations

words related to exclusion or being excluded

oai_token-act-pair · gpt-3.5-turbo

instances of the word "exclude" and its variations, indicating a focus on exclusionary practices or policies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ingham

-0.84

oÄŁ

-0.78

Found

-0.77

 stead

-0.75

nan

-0.69

ebus

-0.67

vous

-0.67

 deed

-0.66

des

-0.66

Ì¶

-0.66

POSITIVE LOGITS

 excluding

0.81

 spoilers

0.78

 bystanders

0.76

ively

0.76

 exclude

0.75

 excluded

0.75

 excludes

0.73

 minded

0.73

 prejudice

0.67

spo

0.64

Activations Density 0.011%

words related to exclusion or being excluded

instances of the word "exclude" and its variations, indicating a focus on exclusionary practices or policies

No Comments

No Known Activations