INDEX

Explanations

phrases related to policies, programs, and organizational initiatives

oai_token-act-pair · gpt-3.5-turbo

terms related to health, safety, and regulatory issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

Laughs

-0.75

Redditor

-0.71

whatever

-0.70

 vomit

-0.70

 booze

-0.69

 Demons

-0.69

fuck

-0.68

 drunk

-0.68

 stupidity

-0.68

Reddit

-0.67

POSITIVE LOGITS

 innovative

0.73

quartered

0.72

 proactive

0.71

 USDA

0.70

 comprehensive

0.69

 federally

0.68

 licens

0.68

 impacted

0.68

 mandated

0.68

 Cosponsors

0.68

Activations Density 1.755%

phrases related to policies, programs, and organizational initiatives

terms related to health, safety, and regulatory issues

No Comments

No Known Activations