INDEX

Explanations

phrases that introduce additional information or context

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

ĸļ

-0.83

fulness

-0.74

INAL

-0.72

 Reincarn

-0.72

 guiActive

-0.72

iker

-0.71

xit

-0.70

arest

-0.70

ibling

-0.69

ħĭ

-0.68

POSITIVE LOGITS

 things

1.42

 places

1.05

 Things

1.05

 thing

1.03

things

0.89

 feats

0.84

 occasions

0.81

 reasons

0.81

 horrors

0.79

 wonders

0.75

Activations Density 0.439%

phrases that introduce additional information or context

No Comments

No Known Activations