INDEX

Explanations

actions or movements described within a narrative context

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

ãĥĩãĤ£

-0.67

ãĤ¦ãĤ¹

-0.61

 extent

-0.59

 sense

-0.59

rists

-0.58

 stakes

-0.57

death

-0.56

 senses

-0.56

anity

-0.56

nov

-0.56

POSITIVE LOGITS

 toward

1.20

 into

1.16

 towards

1.15

 onto

1.13

 ashore

1.07

 hither

1.07

 INTO

1.06

 aboard

1.01

 through

0.98

 away

0.97

Activations Density 1.005%

actions or movements described within a narrative context

No Comments

No Known Activations