INDEX

Explanations

references to the name "Osama bin Laden"

oai_token-act-pair · gpt-3.5-turbo

references to Osama bin Laden

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

anwhile

-0.80

Amazing

-0.74

mble

-0.71

ngth

-0.68

 Parish

-0.68

terday

-0.68

theless

-0.67

ITH

-0.67

 Dragonbound

-0.64

Edited

-0.64

POSITIVE LOGITS

ocular

1.59

omial

1.44

 Laden

1.39

bin

1.00

ning

0.94

ghazi

0.93

utils

0.90

thood

0.89

jamin

0.84

mates

0.82

Activations Density 0.007%

references to the name "Osama bin Laden"

references to Osama bin Laden

No Comments

No Known Activations