INDEX

Explanations

words related to physical locations or structures

oai_token-act-pair · gpt-3.5-turbo

references to hospital wards

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ãĥ¤

-0.77

ãĥĪ

-0.75

istine

-0.75

Hav

-0.70

igslist

-0.69

pheus

-0.69

ctory

-0.68

Bon

-0.67

DonaldTrump

-0.66

issance

-0.66

POSITIVE LOGITS

robe

1.04

 ward

1.03

room

1.02

 wards

0.86

rooms

0.85

masters

0.84

nton

0.83

ring

0.83

stones

0.80

lings

0.79

Activations Density 0.006%

words related to physical locations or structures

references to hospital wards

No Comments

No Known Activations