INDEX

Explanations

references to physical locations or positions, especially focusing on the front or front seats

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

IRO

-0.77

ascript

-0.72

ornings

-0.72

cham

-0.71

cci

-0.71

âĸĪâĸĪâĸĪâĸĪâĸĪâĸĪâĸĪâĸĪ

-0.69

fortune

-0.68

ptives

-0.66

ajor

-0.66

dayName

-0.66

POSITIVE LOGITS

runners

1.08

iers

0.99

 porch

0.95

 lawn

0.92

most

0.84

court

0.83

row

0.83

mast

0.83

 door

0.82

office

0.82

Activations Density 0.398%

references to physical locations or positions, especially focusing on the front or front seats

No Comments

No Known Activations