INDEX

Explanations

words related to negative emotions or events, specifically sadness

oai_token-act-pair · gpt-3.5-turbo

expressions of sadness

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

•layer 3 - sadness

No Comments

Negative Logits

byter

-0.78

iltration

-0.72

orthy

-0.72

Ranked

-0.71

ouver

-0.69

RAFT

-0.67

VERTISEMENT

-0.67

ravings

-0.64

UID

-0.64

IBLE

-0.63

POSITIVE LOGITS

der

1.25

istic

1.08

istically

1.06

omas

1.05

Sad

0.96

sad

0.87

imaru

0.85

die

0.84

stal

0.82

istical

0.76

Activations Density 0.010%

words related to negative emotions or events, specifically sadness

expressions of sadness

No Comments

No Known Activations