INDEX

Explanations

words related to communication, potentially with some emphasis on punctuation

oai_token-act-pair · gpt-3.5-turbo

punctuation and various connectors in writing

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

 encount

-0.74

 cues

-0.71

thora

-0.70

 VIDEOS

-0.70

ias

-0.69

 conduc

-0.69

ussions

-0.67

 bias

-0.67

aid

-0.67

 overboard

-0.66

POSITIVE LOGITS

 Alive

0.82

 Kinder

0.74

liest

0.72

 Offic

0.70

 Literally

0.69

 Anyway

0.69

Ys

0.67

————

0.67

 Stupid

0.66

 Somewhere

0.66

Activations Density 0.898%

words related to communication, potentially with some emphasis on punctuation

punctuation and various connectors in writing

No Comments

No Known Activations