INDEX

Explanations

specific words or phrases related to medical procedures, technologies, and treatments

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

athi

-0.73

MENTS

-0.69

BLE

-0.66

STE

-0.63

Behind

-0.61

atin

-0.59

Bas

-0.59

stad

-0.59

ve

-0.58

Crash

-0.57

POSITIVE LOGITS

 translates

1.15

 means

1.04

 begs

1.03

 explains

1.03

 incidentally

1.01

 resulted

1.01

 implies

0.99

 includes

0.99

 admittedly

0.99

 prompts

0.99

Activations Density 0.099%

specific words or phrases related to medical procedures, technologies, and treatments

No Comments

No Known Activations