INDEX

Explanations

words related to criticizing or evaluating situations

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

dl

-0.68

bridge

-0.67

uble

-0.65

Runner

-0.63

flags

-0.62

ences

-0.61

istor

-0.60

 dismant

-0.60

sil

-0.60

file

-0.60

POSITIVE LOGITS

 considering

0.88

 since

0.74

 pertaining

0.72

 those

0.72

 egregious

0.71

 geared

0.71

given

0.70

 given

0.70

 insofar

0.69

 susceptible

0.67

Activations Density 0.427%

words related to criticizing or evaluating situations

No Comments

No Known Activations