INDEX

Explanations

words related to tools or actions involving a hammer

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

uce

-0.75

RIP

-0.72

VICE

-0.71

ITIES

-0.70

cript

-0.69

ected

-0.69

otes

-0.68

ocard

-0.68

 Charity

-0.67

ceed

-0.66

POSITIVE LOGITS

sonian

1.09

heads

0.93

bats

0.93

head

0.90

 hammer

0.90

hammer

0.90

ing

0.83

lock

0.80

 hamm

0.80

 fists

0.77

Activations Density 0.028%

words related to tools or actions involving a hammer

No Comments

No Known Activations