INDEX

Explanations

information related to news, accidents, incidents, and official statements

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

âĶĢâĶĢâĶĢâĶĢ

-0.71

gets

-0.70

Õ

-0.69

ãĥĸ

-0.68

à¤

-0.68

Aff

-0.67

Translation

-0.67

icative

-0.67

AAAAAAAA

-0.64

AAAA

-0.64

POSITIVE LOGITS

 speculate

0.97

 allege

0.97

hips

0.92

 contend

0.91

paces

0.89

 scrambled

0.88

 consulted

0.87

 believe

0.86

say

0.86

 tasked

0.85

Activations Density 1.051%

information related to news, accidents, incidents, and official statements

No Comments

No Known Activations