INDEX

Explanations

conjunctions 'but,' pronouns 'it' and 'they,' and entities such as 'Red Bull' and 'elected officials' within political and discussion contexts

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

byss

-0.59

htaking

-0.59

 Curious

-0.54

 Trace

-0.51

ministic

-0.50

 Mahjong

-0.49

Pwr

-0.49

Chain

-0.48

Availability

-0.47

clus

-0.47

POSITIVE LOGITS

]"

2.26

],"

2.15

]."

2.13

]'

2.00

],

1.98

]-

1.89

']

1.84

?]

1.79

].

1.78

Activations Density 0.139%

conjunctions 'but,' pronouns 'it' and 'they,' and entities such as 'Red Bull' and 'elected officials' within political and discussion contexts

No Comments

No Known Activations