INDEX

Explanations

phrases related to negotiations or points of contention between parties

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

dl

-0.69

file

-0.67

uble

-0.65

bridge

-0.64

flags

-0.61

inness

-0.60

 dismant

-0.59

figure

-0.58

istor

-0.58

agus

-0.58

POSITIVE LOGITS

 considering

0.82

 those

0.78

 insofar

0.77

 pertaining

0.77

onto

0.72

 targeting

0.71

 ones

0.70

 since

0.70

 geared

0.70

 mentioning

0.69

Activations Density 1.527%

phrases related to negotiations or points of contention between parties

No Comments

No Known Activations