INDEX

Explanations

words related to medical and legal conditions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res_post_32k-oai

Configuration

jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_10.pt

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

32,768

Data Type

torch.float32

Hook Name

blocks.10.hook_resid_post

Hook Layer

Architecture

standard

Context Size

Dataset

Skylion007/openwebtext

Activation Function

topk

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

bright

-0.68

gem

-0.67

ECO

-0.66

 brighter

-0.65

llan

-0.63

irs

-0.63

stand

-0.62

hra

-0.62

gard

-0.61

gaard

-0.61

POSITIVE LOGITS

kefeller

1.23

ution

1.22

utions

1.10

uted

1.04

utes

1.01

uting

0.95

utive

0.94

annabin

0.91

UTION

0.91

uters

0.84

Activations Density 0.110%

words related to medical and legal conditions

No Comments

No Known Activations