INDEX

Explanations

words related to prohibitions or restrictions

oai_token-act-pair · gpt-3.5-turbo

terms related to "prohibition" or restrictions

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

 squared

-0.76

 Primordial

-0.75

 Apostles

-0.67

 rolling

-0.65

 Dover

-0.64

 graves

-0.62

 Span

-0.61

 Spice

-0.60

DOT

-0.60

forth

-0.60

POSITIVE LOGITS

hibit

1.20

hib

1.15

hibition

1.09

oresc

1.04

inous

1.00

itors

0.99

hibited

0.98

itor

0.94

encies

0.94

 exhib

0.90

Activations Density 0.008%

words related to prohibitions or restrictions

terms related to "prohibition" or restrictions

No Comments

No Known Activations