INDEX

Explanations

references to significant geographical entities and their cultural or historical contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Jeremy

-0.76

 playbook

-0.75

 reps

-0.74

 rookie

-0.73

 diapers

-0.73

assion

-0.72

Consumer

-0.71

sych

-0.71

Dispatch

-0.70

 refunds

-0.70

POSITIVE LOGITS

 Located

1.34

Located

1.20

 coastline

1.20

 inhabited

1.19

 inland

1.18

 peninsula

1.18

 southwest

1.14

 mountainous

1.14

 situated

1.13

 UNESCO

1.13

Activations Density 0.458%

references to significant geographical entities and their cultural or historical contexts

No Comments

No Known Activations