INDEX

Explanations

phrases and terms related to the actions and discussions in the United States Congress

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

 Messiah

-0.66

phal

-0.62

OUS

-0.61

other

-0.61

 Templ

-0.61

pul

-0.60

 Prophe

-0.58

 Hedge

-0.58

 Jericho

-0.57

 Worlds

-0.56

POSITIVE LOGITS

woman

1.36

arians

1.19

ional

1.12

men

1.01

crit

0.98

 hearings

0.92

member

0.91

atures

0.91

hips

0.87

members

0.87

Activations Density 0.523%

phrases and terms related to the actions and discussions in the United States Congress

No Comments

No Known Activations