INDEX

Explanations

verbs or phrases related to errors, mistakes, or misjudgments

oai_token-act-pair · gpt-3.5-turbo

terms related to mistakes and errors

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

zona

-0.71

azine

-0.69

icum

-0.68

rise

-0.67

zie

-0.66

rollers

-0.66

properties

-0.65

Jane

-0.65

NetMessage

-0.65

raq

-0.64

POSITIVE LOGITS

 mistakes

0.93

 mistake

0.80

 careless

0.79

 forgiven

0.75

 mishand

0.73

 corrected

0.72

fulness

0.72

 dece

0.71

 prone

0.71

 trusting

0.70

Activations Density 0.052%

verbs or phrases related to errors, mistakes, or misjudgments

terms related to mistakes and errors

No Comments

No Known Activations