INDEX

Explanations

phrases related to measurements and scales, specifically focusing on vertical axes

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

giving

-0.83

unes

-0.82

REDACTED

-0.82

mberg

-0.78

keeper

-0.77

ãģ®éŃĶ

-0.77

erv

-0.76

keeping

-0.74

nil

-0.73

words

-0.73

POSITIVE LOGITS

 axis

0.99

 ascent

0.94

 takeoff

0.88

 cliffs

0.88

 stabil

0.88

 stripes

0.88

 dimension

0.88

 shaft

0.87

 separation

0.84

 leap

0.84

Activations Density 0.019%

phrases related to measurements and scales, specifically focusing on vertical axes

No Comments

No Known Activations