INDEX

Explanations

instances of the word "engage" and its variations related to actions or interactions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res_post_32k-oai

Configuration

jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_6.pt

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

32,768

Data Type

torch.float32

Hook Name

blocks.6.hook_resid_post

Hook Layer

Architecture

standard

Context Size

Dataset

Skylion007/openwebtext

Activation Function

topk

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

NX

-0.66

 Luxem

-0.65

 Cran

-0.64

coat

-0.61

 Walls

-0.61

 Glow

-0.61

 Clover

-0.61

 Pound

-0.60

 Cage

-0.60

 Jorge

-0.59

POSITIVE LOGITS

 meaningful

0.84

 vigorous

0.79

 mutual

0.77

 stren

0.77

 constructive

0.75

 continual

0.74

 reasonable

0.72

 risky

0.71

 mischief

0.69

 intercourse

0.69

Activations Density 0.131%

instances of the word "engage" and its variations related to actions or interactions

No Comments

No Known Activations