INDEX

Explanations

instances where something is given or awarded to someone

oai_token-act-pair · gpt-3.5-turbo

instances of the word "given."

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

ansas

-0.73

ereo

-0.71

itect

-0.71

rians

-0.69

Sphere

-0.69

Phoenix

-0.66

CHA

-0.66

quin

-0.65

atis

-0.64

uder

-0.64

POSITIVE LOGITS

aways

0.90

 precedence

0.87

bread

0.82

away

0.80

 away

0.80

 birth

0.79

 speeches

0.77

 refunds

0.73

 generously

0.73

 sparing

0.73

Activations Density 0.027%

instances where something is given or awarded to someone

instances of the word "given."

No Comments

No Known Activations