INDEX

Explanations

the word "and" in various contexts

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

SHIP

-0.83

udence

-0.79

amily

-0.76

hesda

-0.74

odcast

-0.73

 Administ

-0.72

Deal

-0.71

Report

-0.71

Reviewer

-0.71

Administ

-0.71

POSITIVE LOGITS

 orange

1.13

 yellow

1.13

 striped

1.08

 purple

1.06

 grey

1.06

 stripes

1.04

 blue

1.03

 gray

1.01

 green

0.96

 brown

0.95

Activations Density 0.037%

the word "and" in various contexts

No Comments

No Known Activations