INDEX

Explanations

military-related terms and references

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

 increments

-0.69

Owner

-0.65

yss

-0.64

ASED

-0.62

Transaction

-0.60

 magnitude

-0.58

 bilateral

-0.58

Accessory

-0.57

 implication

-0.57

rawdownloadcloneembedreportprint

-0.57

POSITIVE LOGITS

hips

1.68

hip

1.61

paces

1.44

who

1.39

ervatives

1.30

 alike

1.22

folk

1.19

 whom

1.16

ervative

1.12

Activations Density 1.555%

military-related terms and references

No Comments

No Known Activations