INDEX

Explanations

proper nouns, particularly names of individuals and their associations

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

holders

-0.81

alsa

-0.75

direction

-0.74

 uter

-0.73

illance

-0.72

upuncture

-0.71

ectar

-0.71

exempt

-0.71

planes

-0.69

words

-0.69

POSITIVE LOGITS

 Fischer

1.19

 Reed

1.15

 Robinson

1.13

 Jensen

1.11

Sch

1.11

 Berger

1.10

 Hansen

1.08

 Hayes

1.08

 Petersen

1.08

 Peters

1.07

Activations Density 0.797%

proper nouns, particularly names of individuals and their associations

No Comments

No Known Activations