INDEX

Explanations

words related to congratulations and expressions of support

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

 indo

-0.75

ahime

-0.67

 subsistence

-0.66

atten

-0.64

nit

-0.64

nel

-0.64

²¾

-0.63

ç«

-0.63

 indic

-0.62

eper

-0.62

POSITIVE LOGITS

 congratulate

0.92

ulatory

0.88

him

0.82

 congratulated

0.76

ESCO

0.75

================

0.74

bery

0.74

reon

0.74

worthiness

0.73

kson

0.70

Activations Density 0.008%

words related to congratulations and expressions of support

No Comments

No Known Activations