INDEX

Explanations

phrases indicating acknowledgment or consideration

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Tale

-0.67

 tatt

-0.66

 Pengu

-0.64

 suspic

-0.63

kan

-0.62

 Yose

-0.61

 Courage

-0.60

icht

-0.59

è£

-0.59

 snipp

-0.59

POSITIVE LOGITS

 regards

0.95

 regard

0.74

aning

0.71

reference

0.70

equality

0.70

 permissions

0.69

forward

0.69

 thereto

0.68

ranging

0.68

asers

0.67

Activations Density 0.013%

phrases indicating acknowledgment or consideration

No Comments

No Known Activations