INDEX

Explanations

relationships and dependencies related to quantitative data and variables

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

roo

-0.80

ãĥį

-0.75

umerable

-0.75

ERO

-0.69

ģ«

-0.69

İĭ

-0.69

lyn

-0.65

Ron

-0.65

Appearances

-0.64

Ļ½

-0.64

POSITIVE LOGITS

 luck

0.76

uncture

0.75

 circumstance

0.71

 circumstances

0.70

 severity

0.67

 Advent

0.66

natureconservancy

0.65

luck

0.64

 Unic

0.64

 preference

0.64

Activations Density 0.072%

relationships and dependencies related to quantitative data and variables

No Comments

No Known Activations