INDEX

Explanations

words related to technical details and descriptions, such as model names and product features

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

 mound

-0.67

nikov

-0.66

 square

-0.64

 mull

-0.63

 scar

-0.62

 bloom

-0.61

 lawy

-0.61

 seas

-0.61

 fres

-0.60

laz

-0.58

POSITIVE LOGITS

ERT

0.94

ITION

0.87

 laureate

0.77

NECT

0.76

RD

0.76

arthed

0.76

RN

0.76

XXX

0.75

exempt

0.75

Ni

0.75

Activations Density 2.021%

words related to technical details and descriptions, such as model names and product features

No Comments

No Known Activations