INDEX

Explanations

statements or claims made by characters in dialogue

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

 regress

-0.86

 comparisons

-0.75

 experien

-0.75

anium

-0.74

Reviewer

-0.72

omal

-0.71

 cringe

-0.70

cyclopedia

-0.69

oids

-0.68

etheless

-0.68

POSITIVE LOGITS

 instructed

1.22

 instructions

1.17

 advises

1.09

 commanded

1.07

 advised

1.05

 instruct

1.05

 beck

1.03

Instruct

1.01

 command

0.96

 escort

0.93

Activations Density 1.238%

statements or claims made by characters in dialogue

No Comments

No Known Activations