INDEX

Explanations

terms related to personal situations or experiences, including seeking help or explaining difficulties

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

thood

-0.78

iaries

-0.75

luster

-0.75

ngth

-0.70

oso

-0.68

tackle

-0.68

 CrossRef

-0.67

mage

-0.67

elight

-0.65

airs

-0.64

POSITIVE LOGITS

 rationale

1.24

 reasoning

1.08

 concepts

0.94

why

0.94

 principles

0.92

 virtues

0.91

 criteria

0.90

 workings

0.89

why

0.89

 difference

0.88

Activations Density 0.181%

terms related to personal situations or experiences, including seeking help or explaining difficulties

No Comments

No Known Activations