INDEX

Explanations

references to attributing blame or responsibility

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_post_32k-oai

Configuration

jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_8.pt

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

32,768

Data Type

torch.float32

Hook Name

blocks.8.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

ificent

-0.79

earable

-0.77

enture

-0.77

rocal

-0.76

erial

-0.74

rongh

-0.73

readable

-0.71

oscopic

-0.67

omination

-0.67

estial

-0.67

POSITIVE LOGITS

 faulty

0.90

 blame

0.89

 inaction

0.87

 failings

0.78

 woes

0.77

 inadequ

0.75

 failures

0.74

 fault

0.73

 inexper

0.73

 negligence

0.72

Activations Density 0.789%

references to attributing blame or responsibility

No Comments

No Known Activations