INDEX

Explanations

phrases related to codes of conduct or behaviors that are expected or regulated

oai_token-act-pair · gpt-3.5-turbo

references to conduct or behavior standards and 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

 Dise

-0.67

Cooldown

-0.67

ARK

-0.66

loaded

-0.60

ixed

-0.60

installed

-0.60

 Kers

-0.58

 Lopez

-0.58

iewicz

-0.58

arger

-0.58

POSITIVE LOGITS

onduct

1.22

uations

1.05

ors

0.94

ivity

0.93

avior

0.89

ional

0.88

 conduct

0.88

atform

0.85

ions

0.85

 Conduct

0.84

Activations Density 0.011%

phrases related to codes of conduct or behaviors that are expected or regulated

references to conduct or behavior standards and policies

No Comments

No Known Activations