INDEX

Explanations

names and terms related to individuals or places

oai_token-act-pair · gpt-3.5-turbo

instances of the word "rav" and related names, indicating a focus on violent or destructive behavior

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

ntil

-0.77

ADRA

-0.72

Ŀ

-0.71

ware

-0.70

ĨĴ

-0.68

OPER

-0.67

é¾

-0.65

uador

-0.64

ndum

-0.64

mare

-0.63

POSITIVE LOGITS

iol

0.81

inia

0.79

illac

0.78

inian

0.75

iants

0.73

agement

0.72

iasis

0.72

ashtra

0.71

ion

0.71

iator

0.71

Activations Density 0.061%

names and terms related to individuals or places

instances of the word "rav" and related names, indicating a focus on violent or destructive behavior

No Comments

No Known Activations