INDEX

Explanations

verbs related to communication or action taken by someone

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

theless

-0.78

heid

-0.66

phe

-0.61

urity

-0.61

walker

-0.60

âĶĢâĶĢ

-0.60

vl

-0.60

livious

-0.60

haven

-0.58

arts

-0.57

POSITIVE LOGITS

inel

1.00

ãĥĥãĥī

0.95

keys

0.90

entious

0.90

enced

0.87

 forth

0.80

imental

0.78

 shock

0.74

Grid

0.74

TPS

0.74

Activations Density 0.025%

verbs related to communication or action taken by someone

No Comments

No Known Activations