INDEX

Explanations

phrases related to the potential occurrence or expectation of events

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

 Constant

-0.71

feeding

-0.67

 Palest

-0.64

ament

-0.64

 Fighter

-0.62

shall

-0.60

 Kush

-0.60

 Everywhere

-0.60

 Lifetime

-0.59

checked

-0.58

POSITIVE LOGITS

hap

1.10

haps

1.10

onna

1.10

be

1.07

bes

1.06

 someday

0.94

owe

0.91

ily

0.87

 confuse

0.86

 derive

0.85

Activations Density 0.685%

phrases related to the potential occurrence or expectation of events

No Comments

No Known Activations