INDEX

Explanations

information related to historical events and facts, potentially focusing on statistics and data comparison

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

eny

-0.63

natureconservancy

-0.63

 enthusi

-0.63

past

-0.61

bitious

-0.61

 acceler

-0.60

ierce

-0.59

idespread

-0.59

 masculinity

-0.56

rounder

-0.56

POSITIVE LOGITS

ional

1.24

 maybe

0.85

 insofar

0.81

 occasional

0.77

 perhaps

0.72

 exceptions

0.68

yip

0.67

VPN

0.67

 caveats

0.66

 occasionally

0.66

Activations Density 0.530%

information related to historical events and facts, potentially focusing on statistics and data comparison

No Comments

No Known Activations