INDEX

Explanations

phrases related to serious or harmful actions

oai_token-act-pair · gpt-3.5-turbo

terms related to severe harm or injury

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

XL

-0.72

 Elves

-0.72

fix

-0.71

den

-0.71

wallet

-0.68

girl

-0.67

 Diver

-0.66

gamer

-0.66

cloth

-0.65

starter

-0.65

POSITIVE LOGITS

ous

1.20

ously

1.19

ising

1.18

icates

1.09

ues

1.08

ized

1.08

izations

1.07

icable

1.06

izing

1.06

istic

1.05

Activations Density 0.046%

phrases related to serious or harmful actions

terms related to severe harm or injury

No Comments

No Known Activations