INDEX

Explanations

references to the letter "X" in various contexts

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

-0.71

 Remastered

-0.66

 Rivals

-0.63

 Gree

-0.63

-0.60

 antiquity

-0.60

 alright

-0.57

AVG

-0.57

 Forced

-0.57

 Liberties

-0.57

POSITIVE LOGITS

XXXX

1.08

ibal

1.03

eric

0.91

eros

0.88

ethy

0.86

mas

0.85

ym

0.85

ummy

0.85

ande

0.85

ebra

0.84

Activations Density 0.108%

references to the letter "X" in various contexts

No Comments

No Known Activations