INDEX

Explanations

references to specific locations, possibly related to criminal or political events

oai_token-act-pair · gpt-3.5-turbo

references to specific legal entities and individuals related to court cases

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

poons

-0.76

thia

-0.75

 Corpus

-0.75

piration

-0.71

weights

-0.69

 spaced

-0.69

 corpus

-0.68

 trough

-0.68

creen

-0.67

ffield

-0.67

POSITIVE LOGITS

asonic

0.92

senal

0.86

 Furious

0.81

agascar

0.77

 Express

0.77

 Prin

0.73

icus

0.72

 Fract

0.69

eln

0.69

doms

0.69

Activations Density 0.027%

references to specific locations, possibly related to criminal or political events

references to specific legal entities and individuals related to court cases

No Comments

No Known Activations