INDEX

Explanations

words related to modifying or making changes

oai_token-act-pair · gpt-3.5-turbo

instances of the word "modify" and its variations, indicating a focus on altering or changing something

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.hook_resid_pre

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

çĦ

-0.85

ILLE

-0.76

shows

-0.74

restling

-0.71

ublic

-0.70

Crunch

-0.69

Â¯Â¯

-0.69

Chicago

-0.69

True

-0.68

Â¯Â¯Â¯Â¯

-0.67

POSITIVE LOGITS

atile

0.95

ively

0.80

hap

0.80

 versions

0.75

ations

0.74

 wording

0.74

ives

0.71

 dosage

0.71

iate

0.71

 existing

0.70

Activations Density 0.062%

words related to modifying or making changes

instances of the word "modify" and its variations, indicating a focus on altering or changing something

No Comments

No Known Activations