INDEX

Explanations

instances or references to the word "au"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res_post_32k-oai

Configuration

jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_6.pt

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

32,768

Data Type

torch.float32

Hook Name

blocks.6.hook_resid_post

Hook Layer

Architecture

standard

Context Size

Dataset

Skylion007/openwebtext

Activation Function

topk

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

IONS

-0.67

 Sparkle

-0.67

orative

-0.66

ively

-0.66

ollower

-0.64

 Sagan

-0.62

 Realms

-0.61

bread

-0.60

matically

-0.60

 Grimes

-0.60

POSITIVE LOGITS

llah

1.08

qua

0.94

pt

0.93

gment

0.91

lette

0.91

ppa

0.90

gin

0.86

ction

0.85

cel

0.84

berman

0.83

Activations Density 0.052%

instances or references to the word "au"

No Comments

No Known Activations