INDEX

Explanations

verbs followed by an infinitive 'to'

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Relief

-0.74

Dur

-0.69

orses

-0.68

cedented

-0.66

 Peaks

-0.65

furt

-0.65

 Brach

-0.64

 Starr

-0.64

 teasp

-0.63

EMENT

-0.61

POSITIVE LOGITS

 navigate

1.16

 identify

1.08

 integrate

1.07

 manipulate

1.07

 visualize

1.04

 achieve

1.03

 assemble

1.02

 learn

1.01

 acquire

1.00

 organise

0.99

Activations Density 0.072%

verbs followed by an infinitive 'to'

No Comments

No Known Activations