INDEX

Explanations

mentions of sharp objects like knives and swords

oai_token-act-pair · gpt-3.5-turbo

references to knives and related weaponry

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 buses

-0.71

mberg

-0.71

quickShipAvailable

-0.68

leep

-0.68

ICLE

-0.68

 Revenue

-0.67

anol

-0.67

rians

-0.66

odes

-0.64

urtles

-0.64

POSITIVE LOGITS

 blade

1.25

 blades

1.20

blade

1.18

 wielded

1.14

fish

1.08

 wielding

1.07

tips

1.06

sharp

1.02

 forged

1.01

sonian

1.01

Activations Density 0.086%

mentions of sharp objects like knives and swords

references to knives and related weaponry

No Comments

No Known Activations