INDEX

Explanations

references to specific individuals and their alleged associations or behaviors

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

________________________________________________________________

-0.67

 Vend

-0.63

 loop

-0.62

 Jays

-0.61

 atom

-0.60

OPS

-0.58

 unres

-0.58

 throats

-0.57

 elector

-0.56

 deaf

-0.56

POSITIVE LOGITS

rock

1.01

agher

0.99

berg

0.97

hari

0.94

igan

0.92

son

0.91

enberg

0.91

erer

0.90

enthal

0.89

stein

0.89

Activations Density 0.307%

references to specific individuals and their alleged associations or behaviors

No Comments

No Known Activations