INDEX

Explanations

verb forms indicating action or activity, particularly in the context of roles and functions

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

plet

-0.57

lad

-0.57

ascus

-0.56

 dumps

-0.56

imb

-0.55

cember

-0.55

Jen

-0.53

 Beir

-0.53

TEXTURE

-0.53

Availability

-0.52

POSITIVE LOGITS

ername

1.01

 havoc

0.94

wright

0.92

 roles

0.77

acting

0.76

role

0.76

rored

0.73

offs

0.72

lead

0.72

ulative

0.70

Activations Density 0.034%

verb forms indicating action or activity, particularly in the context of roles and functions

No Comments

No Known Activations