INDEX

Explanations

terms related to counterpart or equivalent entities in different contexts

oai_token-act-pair · gpt-3.5-turbo

references to counterparts in various contexts

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

aping

-0.81

 Butcher

-0.70

haw

-0.69

Cho

-0.68

awar

-0.66

acia

-0.64

pedia

-0.63

redo

-0.63

verbs

-0.63

boat

-0.63

POSITIVE LOGITS

 counterparts

0.97

 counterpart

0.91

ngth

0.77

ries

0.76

MpServer

0.74

sembly

0.73

 abroad

0.73

ican

0.66

awei

0.66

 guiActiveUn

0.65

Activations Density 0.017%

terms related to counterpart or equivalent entities in different contexts

references to counterparts in various contexts

No Comments

No Known Activations