INDEX

Explanations

specific lists or rankings related to character attributes or abilities within a competitive context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-att_32k-oai

Configuration

jbloom/GPT2-Small-OAI-v5-32k-attn-out-SAEs/v5_32k_layer_3

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

32,768

Data Type

torch.float32

Hook Name

blocks.3.hook_attn_out

Hook Layer

Architecture

standard

Context Size

Dataset

Skylion007/openwebtext

Activation Function

topk

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 oneself

-0.82

auri

-0.80

vier

-0.77

 yourselves

-0.72

Tai

-0.69

arov

-0.69

Bush

-0.63

 yourself

-0.62

helm

-0.60

 ourselves

-0.59

POSITIVE LOGITS

 counterparts

1.15

 counterpart

1.01

 selves

0.89

 arsenal

0.86

 brethren

0.84

mates

0.79

 mates

0.73

 holdings

0.73

 portfolio

0.73

 namesake

0.72

Activations Density 9.044%

specific lists or rankings related to character attributes or abilities within a competitive context

No Comments

No Known Activations