INDEX

Explanations

names related to specific regions or individuals

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names of places and characters within a narrative

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

hetical

-0.78

chens

-0.71

ded

-0.70

-0.69

chy

-0.68

ynt

-0.67

amins

-0.66

FML

-0.66

arians

-0.66

pert

-0.65

POSITIVE LOGITS

Gw

1.12

ipeg

0.73

 Savannah

0.72

icago

0.71

Kw

0.69

board

0.68

shield

0.68

aston

0.67

luster

0.67

 Chattanooga

0.67

Activations Density 0.007%

names related to specific regions or individuals

proper nouns, particularly names of places and characters within a narrative

No Comments

No Known Activations