INDEX

Explanations

lists or structured information typically found in instructional or procedural 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

 suspic

-0.78

atan

-0.74

inventoryQuantity

-0.73

soDeliveryDate

-0.72

 warr

-0.71

awaru

-0.69

 psychiat

-0.65

etheless

-0.61

 notor

-0.61

 periodic

-0.60

POSITIVE LOGITS

1.04

:{

0.94

 Recap

0.82

0.80

 onwards

0.80

 Edit

0.80

–

0.80

 Procedure

0.78

:]

0.77

:(

0.76

Activations Density 0.263%

lists or structured information typically found in instructional or procedural contexts

No Comments

No Known Activations