INDEX

Explanations

phrases related to reaching a significant level or point in a process or achievement

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

chance

-0.66

creen

-0.63

 reins

-0.60

few

-0.60

struct

-0.60

 EVENTS

-0.60

currency

-0.59

 sake

-0.59

cell

-0.59

choice

-0.58

POSITIVE LOGITS

 fruition

1.01

 adulthood

0.96

 heights

0.93

 conclusions

0.90

 milestones

0.89

 puberty

0.87

 climax

0.87

 saturation

0.85

 consensus

0.85

 peak

0.84

Activations Density 0.628%

phrases related to reaching a significant level or point in a process or achievement

No Comments

No Known Activations