INDEX

Explanations

words related to jumping or leaping

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

minus

-0.74

 Smile

-0.68

venge

-0.66

ä»£

-0.64

ãĥķãĤ¡

-0.64

ors

-0.64

 reconc

-0.61

nda

-0.61

 detrim

-0.61

ãĥĨãĤ£

-0.61

POSITIVE LOGITS

frog

0.95

started

0.92

start

0.90

lights

0.82

starting

0.82

ulic

0.77

starter

0.77

oons

0.76

erd

0.75

Hop

0.75

Activations Density 0.720%

words related to jumping or leaping

No Comments

No Known Activations