INDEX

Explanations

occurrences of specific names and identifiers in a text, particularly focuses on proper nouns and acronyms

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

çīĪ

-0.74

 Schne

-0.67

 equival

-0.66

 guiActiveUn

-0.64

 Skydragon

-0.62

 guiIcon

-0.61

 circulation

-0.59

 physiology

-0.59

brance

-0.58

 crocod

-0.57

POSITIVE LOGITS

icient

0.91

iqueness

0.77

anamo

0.76

iman

0.76

jan

0.75

ido

0.74

aco

0.74

yo

0.73

imura

0.73

aila

0.71

Activations Density 0.024%

occurrences of specific names and identifiers in a text, particularly focuses on proper nouns and acronyms

No Comments

No Known Activations