INDEX

Explanations

references to actions or events that are expected to happen in the future, often with an intention or prediction associated with them

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

ullah

-0.82

ament

-0.69

aments

-0.68

essor

-0.64

essa

-0.63

 appell

-0.63

 Horus

-0.60

hee

-0.59

clusions

-0.58

lake

-0.57

POSITIVE LOGITS

 overboard

0.88

 downhill

0.80

 nowhere

0.80

 viral

0.78

ggle

0.75

 Commando

0.74

verning

0.73

ãĥ£

0.72

 upstairs

0.72

 forward

0.71

Activations Density 0.166%

references to actions or events that are expected to happen in the future, often with an intention or prediction associated with them

No Comments

No Known Activations