INDEX

Explanations

terms related to being ignored or disregarded in various contexts

oai_token-act-pair · gpt-3.5-turbo

instances of the word "ignore" and its variations, indicating neglect or dismissal of issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

isher

-0.76

uster

-0.72

seed

-0.72

gans

-0.70

ernels

-0.69

iov

-0.69

ood

-0.69

lee

-0.68

amen

-0.67

gins

-0.67

POSITIVE LOGITS

 warnings

1.06

 pleas

0.95

 cues

0.85

 altogether

0.81

 inconvenient

0.81

illy

0.75

 objections

0.74

 complaints

0.73

 responsibility

0.72

 repeated

0.68

Activations Density 0.060%

terms related to being ignored or disregarded in various contexts

instances of the word "ignore" and its variations, indicating neglect or dismissal of issues

No Comments

No Known Activations