INDEX

Explanations

phrases related to negative events or actions that have occurred to individuals in various contexts

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate events occurring after a significant action or circumstance

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

aez

-0.89

tick

-0.75

Mask

-0.73

cin

-0.73

achel

-0.73

Near

-0.73

âĸĪâĸĪ

-0.70

Hyp

-0.67

Í

-0.67

eous

-0.67

POSITIVE LOGITS

 completing

0.88

noon

0.85

market

0.84

 acquiring

0.82

 receiving

0.82

 discovering

0.80

 failing

0.79

 inspecting

0.79

words

0.78

 finishing

0.78

Activations Density 0.111%

phrases related to negative events or actions that have occurred to individuals in various contexts

phrases that indicate events occurring after a significant action or circumstance

No Comments

No Known Activations