INDEX

Explanations

words related to legal or administrative actions and concepts

oai_token-act-pair · gpt-3.5-turbo

past tense verbs and terms related to established entities or systems

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

.","

-0.77

NRS

-0.68

involved

-0.67

SPONSORED

-0.67

 darn

-0.67

-0.64

udeb

-0.61

âĢĳ

-0.60

Ïī

-0.59

OTHER

-0.58

POSITIVE LOGITS

 Features

0.80

 aback

0.79

 Locations

0.77

 Procedures

0.76

 Together

0.76

 Victims

0.75

 Lines

0.75

 Countries

0.74

 Against

0.73

ortunately

0.73

Activations Density 0.152%

words related to legal or administrative actions and concepts

past tense verbs and terms related to established entities or systems

No Comments

No Known Activations