INDEX

Explanations

action-oriented or decision-making words related to various contexts such as finance, sports, and gaming

oai_token-act-pair · gpt-3.5-turbo

references to various roles or participants involved in actions or processes

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

YL

-0.70

REDACTED

-0.68

;;;;;;;;;;;;

-0.66

LOD

-0.64

ascar

-0.63

achment

-0.62

igslist

-0.62

bernatorial

-0.62

 Truth

-0.60

TX

-0.60

POSITIVE LOGITS

 themselves

0.76

mith

0.73

 everywhere

0.71

aurus

0.69

 programmers

0.68

 forgot

0.68

 beware

0.67

 behaved

0.67

poon

0.66

 reverted

0.66

Activations Density 0.308%

action-oriented or decision-making words related to various contexts such as finance, sports, and gaming

references to various roles or participants involved in actions or processes

No Comments

No Known Activations