INDEX

Explanations

terms associated with ambiguity or lack of clarity

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

Negative Logits

 abdom

-0.68

 gorilla

-0.62

 soup

-0.60

 beginnings

-0.59

 sandwiches

-0.59

 antioxid

-0.56

 Âł Âł

-0.56

 specials

-0.56

 Soup

-0.56

 Kingdoms

-0.56

POSITIVE LOGITS

etermined

1.26

oubted

1.24

iscovered

1.17

oubt

1.13

uly

1.11

ifferent

1.10

ef

1.10

ulations

1.08

aunted

1.06

irect

1.05

Activations Density 0.054%

terms associated with ambiguity or lack of clarity

No Known Activations